数据挖掘在阴极结构有限元分析中的应用

Application of data mining to cathode structure design by using finite element analysis

针对场致发射阴极结构设计知识获取问题,提出了采用数据挖掘技术从有限元分析试验结果中提取隐含知识的方案,该方案具有效率高、成本低的特点.分析了包括场致发射有限元分析试验数据结果集获取、数据属性归约等步骤的实现过程,给出了基于熵度量的数据属性集归约和决策树算法原理,从而解决了方案实现的关键问题.以典型结构阴极为例进行有限元分析数据挖掘,通过建立的场致发射性能指标和设计参数关系决策树,提取出了场致发射阴极结构设计知识.算例结果验证了方案的可行性.

Knowledge acquisition was discussed to design cathode structures for field emission. A pro posal with high efficiency and low cost was given, in which implicit knowledge was extracted from the experiment results obtained by finite element analysis （FEA） through data mining technology. The implementation of the acquisition of the experimental data set, the reduction of the data attribution and other steps was analyzed. The principles of data attribution reduction based on entropy measurement and decision tree algorithm were given to solve the key problems in the scheme. Taking a typical cathode structure as an example, the knowledge of designing cathode structure for field emission was acquired by establishing the decision tree of the relationship between field emission performance index and the design parameters. The feasibility of the proposal was verified and theoretical foundation of future experiments was founded by the calculated example.